macOS 26.6.2: Docker Desktop exits and auto-relaunches every ~5m31s after clean reinstall

Open
#643 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Assessment

Difficulty
5/5
Estimated time
Over a week
Newbie friendliness
30/100
Issue type
Bug
Clarity
Mostly clear
Activity status
Active
Tech stack
docker, macos
Domain
desktop

Research direction

Reproduce the report on Apple Silicon macOS 26.6.2 with Docker Desktop 4.88.1, first observing an empty installation and then a Compose stack with running containers. Start with Docker's shutdown logs and the launchd/AppKit events at the reported timestamps, and request the diagnostics bundle from the reporter if needed. Done means identifying the termination trigger and preventing the repeated shutdown and relaunch while containers are running.

Written by the indexing model from the issue text.

Description

Bug report

Description

Docker Desktop repeatedly performs a clean full shutdown and automatically relaunches on macOS 26.6.2. The problem persists after a complete removal of Docker Desktop, its VM disk, containers, images, volumes, settings, CLI configuration, privileged helpers, Homebrew cask, and a clean reinstall of Docker Desktop 4.88.1.

This is not an OOM/Jetsam termination. macOS RunningBoard reports the Docker processes are not memory-managed, and Docker's own log says the virtualization engine shutdown was requested because a termination signal was received.

During the latest clean-install test, five Compose containers were running (including healthy containers), yet shutdown markers recurred at an almost exact 5 minute 31 second cadence:

2026-08-31T12:18:27.325455Z shutdown requested (cancel cause: terminated signal received)
2026-08-31T12:18:27.325505Z backend cancelling with error: <nil>

2026-08-31T12:23:58.818966Z shutdown requested (cancel cause: terminated signal received)
2026-08-31T12:23:58.824735Z backend cancelling with error: <nil>

2026-08-31T12:29:30.204677Z backend cancelling with error: <nil>
2026-08-31T12:29:30.205254Z shutdown requested (cancel cause: terminated signal received)

The first observed sequence shows:

  1. Docker logs terminated signal received at 12:18:27.325Z.
  2. A later teardown-side AppKit fault appears at 12:18:27.420Z:
    Unrecognized event type 0 / NSInternalInconsistencyException.
  3. launchd reports the top-level Docker application exiting normally with exit(0) at 12:18:29.775Z.
  4. Docker Desktop automatically relaunches afterward.

The AppKit fault occurs after shutdown has already started, so it appears secondary rather than the initiating cause.

The workaround discussed in #434 was also tested before the clean reinstall (UseResourceSaver=false, AutoPauseTimeoutSeconds=0), followed by a full stop/start. It did not prevent the termination. No docker mcp gateway run process was present.

Reproduction steps
  1. Install Docker Desktop 4.88.1 on Apple Silicon macOS 26.6.2.
  2. Start Docker Desktop.
  3. Optionally start a Compose stack; the reproduced run had five running containers.
  4. Observe Docker Desktop for several minutes.
  5. Docker Desktop shuts down cleanly, the engine and running containers disappear temporarily, and the application automatically relaunches.
  6. The behavior repeats at approximately 5 minute 31 second intervals.
Expected behavior

Docker Desktop and its engine remain running until explicitly stopped, especially while containers are running.

Troubleshooting already performed
  • Verified the Docker application signature with codesign --verify --deep --strict.
  • Completely removed Docker Desktop and all Docker runtime data, including Docker.raw, settings, containers, images, volumes, CLI configuration, app bundle, Homebrew cask, privileged helpers, launch daemons, and Desktop-created command links.
  • Reinstalled the current Homebrew cask and ran Docker's privileged helper installer.
  • Confirmed the fresh engine started empty before a separate Compose workflow populated it.
  • Confirmed this is not host memory pressure or Jetsam.
  • Checked user/system LaunchAgents, LaunchDaemons, and crontab for Docker kill/quit automation; none was found.
  • Confirmed no Docker MCP gateway process was running.
  • The issue reproduced under load with running containers.

Platform

macOS

Version information

macOS 26.6.2 (Build 25G83)
Apple M4 Pro
arm64

Docker Desktop 4.88.1 (237512)

Client:
 Version:           29.7.2
 API version:       1.55
 Go version:        go1.26.5
 Git commit:        a7dcaa6
 OS/Arch:           darwin/arm64
 Context:           desktop-linux

Server: Docker Desktop 4.88.1 (237512)
 Engine:
  Version:          29.7.2
  API version:      1.55 (minimum version 1.40)
  Go version:       go1.26.5
  Git commit:       6a43e3d
  OS/Arch:          linux/arm64
  Experimental:     false
 containerd:
  Version:          v2.3.3
  GitCommit:        aad11006b869517fcd3009450b6f82da282e1a9b
 runc:
  Version:          1.4.3
  GitCommit:        v1.4.3-0-gbb14dabe
 docker-init:
  Version:          0.19.0
  GitCommit:        de40ad0

Diagnostics ID

Not uploaded because the full diagnostics bundle may contain sensitive local project and container details. I can provide a diagnostics ID privately to Docker maintainers if requested.

Dominant language
No language data
Stars
46
Forks
20
PR merge metrics
No merged PRs in 30d

Contributor guide

No contributing guide indexed for this repository

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

More from docker/desktop-feedback

All issues in docker/desktop-feedback

Similar issues

More Desktop Dev issues

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.